So I’ve been playing around with Dataverse a little more today in a continuation of yesterday’s fiasco! So if you haven’t already seen my previous post, best start by reviewing:
Below are my preliminary observations and my impressions may change with time and experience, but as of today, this was my impression of things.
The Good
So, I finally managed to import a table and its records.
In my very preliminary testing, data read/write speeds were quite good. This of course was under light load. Much more testing is required.
Tomorrow, I will be attempting to import more than 1 table with relationships, again. Hopefully it will go better than it did in my first attempt!
The Bad
The import process is horribly slow. I received several “The application is not responding.” error messages.
Personally, I find the Dataverse web UI to be very crappy! I suspect I’ll be having lots of fun working with it in the near future. Add to that, the manner in which Microsoft buried/hid the Dataverse admin page just blows the mind. Microsoft couldn’t make it harder for someone to get to it, even if they tried!
When a local table is imported and then linked, any relationships are not recreated between other local tables and the linked table. ***** Upon my forth attempt, the relationships were recreated. So it didn’t work and this time it did?! I don’t know what is going on. ****
When imported the original local table is renamed. This renamed table retains all its relationships so be forewarned as this can cause issue since you probably want the linked Dataverse table to be the main table, not this backup copy.
One things that is causing me some pain is the fact that when the table is created in Dataverse, it sorts the fields/columns alphabetically. So it reorders the fields/columns. Therefore, you loose any organization you may have created, … this can cause unnecessary headaches especially if you are trying to compare the original table, and/or relationships, with new Dataverse versions.
Also note, that any Dataverse Web UI work is slow. Navigating the PowerApps site is complex and confusing initially and everything is on a different page so you’re loading over and over and over. There is simply no comparison to local Access development.
Links:
To try and save you the same pain I feel everytime I try to find anything in the maze of menu on the Power Platform site, … here is/are some direct link(s):
Environments -> https://admin.powerplatform.microsoft.com/environments/
The Ugly
As mentioned in part 1, the import process appears to be VERY unreliable!
Trying to share a Dataverse driven Access database with ‘guests’ (my client) is a NIGHTMARE! Heck, PowerApps, Azure, … administration is a complete nightmare!!! I even asked the Access Dev Team and they couldn’t offer any guidance on the matter. Their product uses a technology, but they can’t help in what would be expected normal usage. So you develop a solution for a client, and have no way to provide it to them for demonstration purposes.
I’m also VERY concerned about whether I can even migrate tables from a development account (ie mine) to a production account (ie my client’s). I suspect I may be in for a real problem here!
Where I Stand On The Subject Today
Right now, with the headaches I’ve encountered, problems I’m still facing (simply trying to share a database with a client) and future unknowns (needing to deploy the Dataverse tables to another account), I rate this whole Access Dataverse thing a 3 out of 10.
For the average developer working to create internal tools, not having to worry about external users, deployment, … I’d probably push my rating to a 4-5 out of 10.
Someone starting with a proper database and that does truly proper development (so experienced developers) for internal usage only … we could push that to a 6-7 out of 10.
My biggest takeaway from this whole experience has been that you’ll need to arm yourself with a LOT of patience! And if you’re a religious person, a lot of prayers!
Also, I am still hopefully, given time, Microsoft will iron out the issues I experienced.
I should further mention that you should be aware that currently Dataverse cannot import/handle all of Access’ data type, so that alone can be a deal breaker in some instance. I also know, that they are working on expanding the data types that are accepted and no doubt eventually I would think all data types will work, let’s hope.
More to come … eventually … if I decide to keep fighting with Dataverse.
Lastly, personally, unless something drastically changes, even if Dataverse worked flawlessly, I truly don’t see the advantage it provides and would simply use an Azure SQL database (if you want to stick with a Microsoft solution). This would put me in the driver seat with full control of the database, unlike Dataverse where you can’t manage things (or are very limited in your ability to do so). Remember Dataverse is in fact built upon an Azure SQL database. Think long and hard before deciding to take the Dataverse dive!
Dataverse/Powerapps is another LightSwitch waiting to be switch off. If their is a project or software that will convert Access ACCDB or ACCDE into Web/Mobile App, that would be a NOBEL PRIZE! TB is trying to clone Access! hopefully. Thanks Dan for being so generous.
“Dataverse/Powerapps is another LightSwitch waiting to be switch off.”
Previous attempts to ‘webify’ Access I’d agree, but Microsoft has invested hugely into Dataverse, I don’t think it will be going anywhere anytime soon. After all, Dataverse is built upon Azure SQL.
“If their is a project or software that will convert Access ACCDB or ACCDE into Web/Mobile App, that would be a NOBEL PRIZE”
None exist that I’m aware of, but as I’ve stated, it can still be done. It take some work, but in the end the results, IMHO, are well worth it.
“TB is trying to clone Access!”
TB? Twin Basic?
I’ve always been amazed that no other major IT player has taken a serious stab at Access. Oh, there are a multitude of databases, but not a designer like Access.
“Previous attempts to ‘webify’ Access”. To my recollection, if only MS improved-enhanced and did their focus on the Data Access Page technology which was introduced in Access 2000 then the Access WEBFICATION I think was a success. The failed Access WEBFICATION was intentional in nature or else Access will eat most of the web DB centric tolls inside MS. Imagine if you can produce a responsive template based web app from DAP? it’s game over!
>> I don’t think it will be going anywhere anytime soon. <<
In other words, it's going nowhere! 😉
d
Ouch!
Bring back QA – its not like Microsoft don’t have any spare cash.
That ship has sailed, sadly. Microsoft has gotten every accustomed to routine bugs and no one care now. Everyone just puts up with the continuous monthly bug release by Microsoft. The time to do something was when Office 2016/Office 365 first came out. If people had ditched Office then, Microsoft would have reacted, but now, years later, with no one truly having done anything and record profits you can be sure they aren’t going to do that. Not a chance in hell! It’s all about the $$$.
If Microsoft are not careful everyone will gradually migrate to something like cpanel on linux with bog standard mysql databases.
I’ve started hearing about Internal Development Platforms all over the place. The new name for pretty familiar concept..
Nothing wrong with using MySQL to build apps off of, I’ve done so myself and it works flawlessly! (if done properly!!!)
Oh dear. I was very excited when I first hear about Dataverse and Access “working together”. I just ready both your articles and now I am no longer excited.
Thank you for your efforts testing and giving nthis feedback. One can only hope that one day, it actually works.
I started learning Power Apps a month ago and developed a couple of apps, and it wasn’t the most terrible experience in the world. I came from traditional Access & VBA environment just like Daniel and most of you, but I don’t feel as negatively about the Power platform as Daniel does. And I don’t know why he had such a hard time doing a simple thing like importing a table, which was a cinch for me (I did it from within Power Apps, so maybe that was the difference?).
Here is a video demo of a simple movie collection app I made with Power Apps in a couple of days. It shows movie info, cast and crew, cover images, etc.: https://www.youtube.com/watch?v=xq5cs4_h9hc
The biggest hurdles I’ve found so far are two things. First is something called “delegation”. Some tasks in Power Apps can be “delegated” to your data storage, meaning the tasks can be assigned normally to the engine of your data source (Dataverse, Sharepoint, or SQL Server) with no problem, thus lightening the load on Power Apps. But if a task can’t be delegated, Power Apps will have to do the task itself; as a result of the increased load, Power Apps will then only process up to 2000 records from any table in your data source. This will potentially ruin any operations that involve row count or row summation. And if you app can’t even count or sum correctly, your app is just no good. So this “delegation” issue is a pretty big deal, and app designers should make it top priority to learn about delegation and how to avoid “non-delegable” tasks.
The second hurdle (a relatively minor issue) is that there is no built-in way to make an input mask for a text box like we can in Access. It’s a trivial thing to do it in Access. But in Power Apps, if you want a simple input mask for phone numbers like “(000) 000-0000”, you need to write some heart-stoppingly complicated code.
Thanks for the write-up, Daniel – very informative.
I need to disagree with one of your Ugly points though. I would not expect tables in two different databases (Dataverse and local tables) to be able to enforce referential integrity. That’s not possible between SQL and Access, so why should it be here?
You are entirely right Armen! It had been a long couple of days fighting with DataVerse when I wrote the post. I will revise it.